Signature-based IaaS Performance Change Detection

2024-10-23

ACM Transactions on Internet Technology

AI Summary

This paper addresses the challenge of distinguishing stochastic noise from genuine performance shifts in long-term IaaS service monitoring. To this end, we propose an automated mutation detection method based on performance signatures. Our approach introduces a lightweight performance signature representation model that integrates sliding-window analysis with time-series similarity metrics. Crucially, we establish— for the first time—theoretical IaaS performance noise modeling and an SNR-driven change discrimination mechanism, enabling robust decoupling of noise from true performance drift. Extensive experiments on real-world IaaS datasets demonstrate that our method significantly improves detection accuracy and reduces false positive rates, outperforming state-of-the-art baseline approaches across multiple metrics. The framework delivers interpretable, production-ready support for sustainable cloud infrastructure performance governance.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Evaluation and Analysis
• Search and Optimization: Evaluation and Analysis

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Web Mining and Content Analysis: Models for Web evolution
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

We propose a novel change detection framework to identify changes in the long-term performance behavior of an IaaS service. An IaaS service’s long-term performance behavior is represented by an IaaS performance signature. The proposed framework leverages time series similarity measures and a sliding window technique to detect changes in IaaS performance signatures. We introduce a new IaaS performance noise model that enables the proposed framework to distinguish between performance noise and actual changes in performance. The proposed framework utilizes a novel Signal-to-Noise Ratio (SNR) based approach to detect changes when prior knowledge about performance noise is available. A set of experiments is conducted using real-world datasets to demonstrate the effectiveness of the proposed change detection framework.

Problem

• Detect long-term IaaS performance changes
• Distinguish performance noise from actual changes
• Utilize SNR for change detection with noise knowledge

Innovation

• Time series similarity measures
• Sliding window technique
• SNR-based change detection